Most US companies lack mature AI governance frameworks as agentic AI spreads

Most US companies are deploying AI agents faster than they can govern them. A new Schellman report reveals that while 86% of organizations are testing AI agents and nearly half have them in production, only 64% have a formal AI acceptable use policy actively communicated to employees. For AI builders, this governance gap is both a risk and an opportunity: companies with mature AI governance frameworks report easier scaling, higher customer trust, and better regulatory readiness.

What happened

Agentic AI is spreading rapidly across US organizations. According to the Schellman report, 86% of organizations are testing AI agents and nearly half already have them in production. Yet governance has not kept pace. Only 64% have a formal AI acceptable use policy that they actively communicate to employees, 57% maintain formal policies, and just 44% have AI-specific incident response procedures.

A separate IBM Institute for Business Value study found that two-thirds of CIOs and CTOs say they are held accountable for AI systems that they do not fully control. A Deloitte study from earlier this year found that only 1 in 5 companies have a mature model for governing autonomous agents.

Why AI builders should care

Governance is shifting from optional to mandatory. Avani Desai, CEO of Schellman, said in a statement that customers, regulators, boards, and business partners want proof that governance is working. A majority of companies reported preparing to comply with US regulations, and nearly one-third are preparing to comply with the EU AI Act.

For AI builders, this means that enterprise buyers will increasingly require evidence of governance maturity before adopting your product. If you are building AI agents or tools for enterprise use, your customers will expect you to help them meet governance requirements around acceptable use, incident response, and accountability.

Practical implications

Organizations that deploy meaningful governance see tangible value. Respondents in the Schellman report said they were seeing improved internal efficiency, an easier time scaling AI and innovation, and an increase in customer trust.

Governance should not be treated as a one-time exercise. As companies deploy more autonomous AI capabilities, they need continuous processes of oversight, accountability, and validation. For AI builders, this suggests an opportunity to build governance tooling that integrates into existing workflows rather than requiring separate audits.

Caveats

The data in this article comes from industry reports and surveys, which rely on self-reported responses. Concrete company-level data on governance maturity varies by source. The Schellman report, IBM study, and Deloitte study each use different methodologies and sample sizes, so exact percentages should be treated as directional rather than precise. Additionally, the report does not specify which industries or company sizes are most affected, so your mileage may vary depending on your target market.